1. the
science
of
marketing
Sampl
Excer e
p
Only t
when to tweet, what to post,
how to blog, and other
proven strategies
Dan ZarreLLa
2. WHEN TO TWEET, WHAT TO POST,
HOW TO BLOG, AND OTHER
PROVEN STRATEGIES
3. 3 how to attract customers with twitter & Vine
HubSpot’s All-in-One
M w
Email SOCIAL
Marketing Software MARKETING CONTACTS
HubSpot brings your whole marketing
world together in one, powerful, integrated
system.
Get Found: Help prospects find you online
Convert: Nurture your leads and drive conversions
U
CONTACTS
DATABASE
;
WORKFLOWS
Analyze: Measure and improve your marketing.
More: See marketplace for apps and integrations
Request A Demo Video Overview
n
SMART FIELDS analytics
Share This Ebook!
www.Hubspot.com
4. Contents
Acknowledgements ix
Introduction 1
Part I Content 7
Chapter 1 E-Books 9
Chapter 2 Webinars 25
Part II Channels 35
Chapter 3 SEO 37
Chapter 4 Twitter 53
Chapter 5 Facebook 75
Chapter 6 Pinterest 101
Chapter 7 Blogging 109
vii
5. viii Contents
Part III Middle of the Funnel (MOFU) 127
Chapter 8 E-Mail Marketing 129
Chapter 9 Lead Generation 153
Part IV Analytics 171
Chapter 10 Analytics 173
Index 185
7. Twitter is my favorite social site. I love the simplicity, the flex-
ibility, and the vast audience. I remember a time before the word
retweet existed, when it took only 30 or so tweets from about as
many people for a phrase to become a trending topic worldwide.
It is the perfect platform for the distribution of marketing content.
Describe a link, paste it in the box, and hit Tweet. Your followers
can then click and read, and if they’re so motivated, they can share
that link with their followers. It’s the most elegant viral mechanism
yet invented.
I hold some controversial points of view about Twitter, but
none without data backing them up. And that’s what this chapter
is—my most important Twitter data (and the best collection of it
anywhere).
I’ve long been interested in the idea that “engaging in the con-
versation” is the single most important function of social media
marketing, so I’ve applied my analysis to test that statement in a
variety of places. One of those places has been Twitter.
I looked at millions of Twitter accounts and separated them into
two groups: those with more than 1,000 followers (the first orange
bar in Figure 4.1) and those with fewer than 1,000 followers (the
first black bar in Figure 4.1). I then compared those two groups by
the percentage of their tweets that started with an “@” sign to arrive
at a reply percentage. I repeated this analysis with accounts having
more than 1 million followers (the second orange bar in Figure 4.1)
and accounts with fewer than 1 million followers (the second black
bar in Figure 4.1) and found similar results.
55
8. 56 Channels
18%
Percentage of Tweets That Start with “ @ ”
16%
14%
12%
10%
8%
6%
4%
2%
0%
>1,000 <1,000 >1 million <1 million
Figure 4.1 Reply Percentage and Follower Count
50%
Link and Nonreply Percentage
40%
30%
20%
10%
0%
>1,000 <1,000
Figure 4.2 Highly Followed Accounts Tweet Lots of Links
Highly followed accounts tend to spend a lower percentage of their
tweets replying to other accounts—they are less conversational—
than less followed accounts.
The first question I thought when I uncovered the data in Figure
4.1 was, “If they’re not engaging in the conversation, what are they
doing?” So I did the same breakdown of more than 1,000 versus
fewer 1,000 followers—but this time I analyzed the percentage of
tweets that did not start with an “@” sign and that contained a link
(Figure 4.2). I measured how much content was being broadcast by
these accounts.
9. Twitter 57
I found that highly followed accounts tweet more links
than their lesser followed counterparts. These accounts did not
build their reach by being in conversations; they built it by sharing
interesting content in a broadcast fashion. In fact, there are not
many examples of well-known Twitter accounts that are built on
lots of replies, whereas there are countless accounts with more than
1 million followers that do nothing more than share interesting
facts, quotes, links, and news.
Do not think of “engaging in the conversation” on Twitter as a
way of building your reach. Instead, focus on gathering and sharing
as much interesting, relevant content as you can.
If you should tweet lots of links to get followers, how many links
is too much? Is it possible to overtweet?
Using data from HubSpot’s free Twitter Grader tool, I analyzed
just over 5 million Twitter accounts and compared the number of
times per day they tweeted on average and their number of follow-
ers (Figure 4.3). I found that followers peaked with accounts that
1,200 1,000,000
R2=0.6627
1,000
100,000
800
600 10,000
400
1,000
200
0 100
1
7
13
19
25
31
37
43
49
55
61
67
73
79
85
91
97
103
109
115
121
127
133
139
145
Tweets per Day
www.Hubspot.com
Figure 4.3 Tweets-per-Day versus Followers
10. 58 Channels
tweeted around 22 times per day and there was no steep drop off
beyond that.
Twenty-two tweets per day, on average, is a pretty breakneck
pace for most accounts to keep up, especially if they’re tweeting
interesting content and not just anything they find. The takeaway
of this graph is not that 22 times a day is a magical number, just that
it’s pretty hard to overtweet. And if you’re wondering how often you
should tweet, the answer is generally “more than you currently are.”
During my research into Twitter, I translated two linguistic anal-
ysis systems to the microblogging platform: Linguistic Inquiry and
Word Count (LIWC) and Regressive Imagery Dictionary (RID). I
did not invent these two systems; they were created by academics at
universities. I simply applied them to social media.
One of the traits that these systems allowed me to analyze was
self-referential language—how often accounts refer to themselves,
either as individuals or as an organization. This includes use of
words such as I, me, us, and we.
When I compared the percentage of tweets that used self-
referential language to the number of followers those accounts
had—looking at millions of accounts—I found a striking pattern
(Figure 4.4). As self-reference increases, follower count decreases.
1.3%
1.2%
Self-Reference
1.1%
1.0%
0.9%
0.8%
0.7%
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
Followers
Figure 4.4 Effect of Self-Reference on Followers
11. Twitter 59
For most noncelebrity individuals and brands, Twitter users do not
follow them to hear them talk about themselves constantly.
Those same linguistic analysis systems also allowed me to look
into the relationship between negative sentiment and followers.
Here also, I found a similar, striking pattern. As negative remarks
increase, follower counts decrease (Figure 4.5).
People don’t go to social media to get bummed out about the
world around them; they can just turn on the TV news if that’s
what they want. They go to social media to talk to their friends and
generally feel good.
This sounds like unicorns-and-rainbows superstition, but in this
case, the data support it. Negativity doesn’t sell on social media as
well as positivity does.
When you sign up to Twitter, you’re given the ability to provide
three bits of personal information: a profile picture, a 160-character
bio, and a link to your home page. Over the years I’ve tracked the
number of accounts that fail to fill these fields out, and although
the numbers have gotten better with time, I’m still surprised by how
many accounts don’t take the few moments required to do this.
When I analyze the relationship between providing this infor-
mation and follower counts, the results are unsurprising. In all
three cases, accounts that provide a picture, bio, and home page
1.0%
0.9%
Negative Remarks
0.8%
0.7%
0.6%
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
Followers
Figure 4.5 Effect of Negative Remarks on Followers
12. 60 Channels
300
Average # of Followers
200
100
0
Picture No picture
Figure 4.6 Effect of Profile Picture on Followers
link all have more followers than accounts that do not. Figure
4.6 shows the number of followers versus having or not having a
profile picture, but the effect is the same with the other two fields
as well.
Take the time to fill out your Twitter bio. Users want to know
who you are before they’ll follow you.
Going a step further, I dug into actual language used by Twitter
account holders in their bios. One common unicorns-and-rainbows
myth is that you should not call yourself a guru or use any other
word to label yourself an expert.
But the data contradict this superstition. I found that Twitter
accounts that used the word guru in their bios had about 100 more
followers than the average account (Figure 4.7). These data do not
mean that if you go over onto Twitter right now and add that word
to your profile that you’ll instantly get more followers. But if you
look at the rest of the graph, it does indicate that you should not
be afraid to identify yourself authoritatively. Tell potential followers
why they should listen to you. If you’ve written a book, founded a
company, or are an expert on something, tell us.
But remember the data about self-reference in tweets. The bio is
the only place you should be talking about yourself.
13. Twitter 61
200
Followers above Average
150
100
50
0
Official
Founder
Speaker
Expert
Guru
Author
Figure 4.7 Effect of Bio Words on Followers
Sweet spot
16%
Percentage of Tweets Retweeted
14%
12%
10%
8%
6%
4%
2%
Percentage of Tweets Containing a Link
0%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Figure 4.8 Sweet Spot for Retweets: Links
So far, I’ve presented data only about follower counts, but
I’ve also spent a great deal of time analyzing retweets as well. In
Figure 4.8, I compared the percentage of accounts’ tweets that con-
tain a link and their average number of retweets per tweet.
I found that here is a sweet spot of linking for maximum retweets.
Accounts that posted 60 to 80 percent links tended to get the most
retweets. Ninety percent or more links can look spammy, so retweet
performance tends to drop off there.
14. 62 Channels
Sweet Spot
Percentage of Tweets Retweeted 16%
14%
12%
10%
8%
6%
4%
2%
Percentage of Tweets Beginning with @
0%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Figure 4.9 Sweet Spot for Retweets: Replies
Then I looked at a similar comparison of reply percentage
and retweets and found a much different sweet spot (Figure 4.9).
Basically, the more replying done by the Twitter accounts in my
data set, the lower their average retweets per tweet.
As these data and the data about follower counts show, con-
stantly replying and being chatty on Twitter does not benefit a mar-
keter in terms of reach or content spread. If you’re using Twitter
as a marketing channel, with the goal of building a large audience
of engaged followers who often share your content, you’ll be best
served by focusing on sharing a lot of interesting content, rather
than replying to every message you get.
When I first began my retweet research, the easiest way I found
to get more retweets was to simply ask for them (Figure 4.10). In
fact I conducted an experiment before the word retweet had been
invented where I simply asked people to tweet a certain link, and
the title of the experiment ended up trending worldwide.
Over the years since then, I’ve had many discussions and debates
about the power of asking for retweets. More recently, I decided to
update and solidify my data about it to put an end to the doubt.
I looked at 20,000 randomly selected tweets and broke them
into three groups: those that contained the phrase please retweet,
those that contained please RT, and those that contained neither
call to action. I found that whereas only 12 percent of the “neither”
15. Twitter 63
0% 10% 20% 30% 40% 50%
Please Retweet 51% RT’ed ± 1.69%
Please RT 39% RT’ed ± 1.43%
Neither 12% RT’ed ± 0.67%
Confidence interval: 90%
Figure 4.10 Phrase Please Retweet Gets Four Times More Retweets
group’s tweets were retweeted, more than 50 percent of the “please
retweet” group’s tweets were.
Calls to action work in all forms of marketing, and social media
is no different. If you want more retweets, ask for them.
Thanks to the great people over at Buffer (an awesome app you
should check out), I was able to study millions of tweets and their
retweet performance. First, I looked at the relationship between the
time of day the tweets were sent and how many retweets they got
(Figure 4.11). The data confirmed my earlier analysis on a different
data set and showed that retweets were highest for tweets posted
between 3 pm and 5 pm Eastern time.
I know from personal experience that as the business day wears
on, I often lose the motivation and wit to come up with worth-
while original tweets. It’s around 4 pm that my retweeting activity
increases because of this.
Experiment with tweeting those updates you want to spread
during this time period and see if it works for you.
The first large-scale retweet data set I compiled was more than
100 million retweets gathered over the course of more than a year.
It was this database that formed the basis of much of my earliest
work on Twitter.
Using this data set, I was able to analyze the volume of retweet-
ing activity that occurred on the different days of the week. I found
that although overall Twitter activity tends to be highest early in
the business week, retweeting peaks on Fridays (Figure 4.12).
16. 64 Channels
6
5
4
3
2
1
0
12:00 am
1:00 am
2:00 am
3:00 am
4:00 am
5:00 am
6:00 am
7:00 am
8:00 am
9:00 am
10:00 am
11:00 am
12:00 pm
1:00 pm
2:00 pm
3:00 pm
4:00 pm
5:00 pm
6:00 pm
7:00 pm
8:00 pm
9:00 pm
10:00 pm
11:00 pm
Figure 4.11 Time of Day versus Retweets
I think the reasons for this might be similar to the reasons that
the end of the business day is the best for getting retweets. And
the tactical takeaway is much the same as well. On Friday after-
noons, look over the content you’ve posted during the week and
share the best stuff again. After a few weeks, you’ll learn if Friday
afternoon as a highly retweetable time works for your business and
your audience.
Earlier in this chapter, I included data that showed that self-
reference was correlated with lowered follower counts. I also ana-
lyzed the relationship between self-reference and retweets and
found a similar pattern (Figure 4.13).
Looking at millions of retweets and millions of non-retweeted
“normal” tweets, I found that non-retweets tended to contain more
self-referential language than retweets. Not only does self-reference
not lead to more followers, it also doesn’t lead to more retweets.
When I share one of your tweets with my audience through a
retweet, I need to believe that it will be relevant and interesting to
my followers. The minutia of your life—what you had for lunch and
how many times your cat farted—is very unlikely to make that cut.
If you’re on Twitter to communicate with friends you know in real
17. Twitter 65
19%
Retweet Volume 17%
15%
13%
11%
Sun Mon Tues Wed Thu Fri Sat
Figure 4.12 Retweet Activity by Day
3%
2%
1%
0%
Retweets Non-retweets
Figure 4.13 Amount of Self-Reference in Tweets and Retweets
100
80
Word Occurrence
60
40
20
0
Retweets Random tweets
Figure 4.14 Word Novelty in Retweets
18. 66 Channels
life, feel free to talk about yourself all day long, but if you’re there
for marketing and business reasons, stop talking about yourself.
Whenever I’ve asked people, in surveys or focus groups, why
they retweet some tweets but not others, the idea of novelty comes
up frequently. People tell me that they want to retweet new infor-
mation. They want to be the first, not the last, to inform their
followers of some breaking news. Scarce and new information is
valuable information. Things that everyone else knows aren’t par-
ticularly worthwhile.
In an effort to quantify this idea, I looked at word occurrence in
retweets and non-retweeted normal tweets. I measured how com-
mon the words in each tweet were (the is a very common word and
thus had a high word occurrence score, whereas sesquipedalian is
much less common and has a very low word occurrence score).
I found that retweets tended to contain rarer words than non-
retweeted tweets. Nobody wants to retweet you if you’re simply say-
ing the same things everyone else is saying. If you want me to share
your content, you need to say something new, something I (and my
followers) haven’t heard—or read—before.
Perhaps the simplest bit of retweet analysis I’ve conducted is
about the occurrence of links in tweets and the likelihood that
those tweets are retweeted. In my data set, I found that only 18.96
percent of tweets contained a link, but 56.69 percent of retweeted
tweets contained a link. People are more likely to retweet a link
rather than just a simple tweet (Figure 4.15).
These data should serve to reinforce the importance of sharing
as many interesting, relevant links on Twitter as you can. Share
your content—and don’t be afraid to share it a few times—and find
content from other sources that will also interest your audience and
share those links. Establishing yourself as a source of useful, novel
content is the most data-supported strategy to more followers and
retweets.
I then analyzed the most common words and phrases that
occurred in retweets more than they’re expected to, based on how
often they occur in non-retweeted normal tweets (Figure 4.16). I
found a number of interesting things, but we should remember to
19. Twitter 67
Non-retweets Retweets
Links
Links
Figure 4.15 Effect of Links on Retweets
1. you 11. please retweet
2. twitter 12. great
3. please 13. social media
4. retweet 14. 10
5. post 15. follow
6. blog 16. how to
7. social 17. top
8. free 18. blog post
9. media 19. check out
10. help 20. new blog post
Figure 4.16 Most Retweetable Words and Phrases
think about the reasons why these words are on this list, rather than
blindly using them and expecting more retweets.
The most retweetable word in my data set was the word you.
Twitter users want to hear you talk about them, not yourself. The
words twitter and social indicate that talking about social media in
general and Twitter in specific works on Twitter. Also on the list
are please retweet and new blog post, which corroborate earlier points
about asking for retweets and the importance of novelty. Free is
always a powerful word in marketing, and on Twitter it is no differ-
ent. And we find how to, top, and 10 on this list, showing that utility
content and chunked, list-based content performs well on Twitter,
as it does on other forms of social media.
On the flip side of the coin, we find the least retweetable words,
those words that occur far less in retweets than their commonality
20. 68 Channels
1. game 11. well
2. going 12. sleep
3. haha 13. gonna
4. lol 14. hey
5. but 15. tomorrow
6. watching 16. tired
7. work 17. some
8. home 18. back
9. night 19. bored
10. bed 20. listening
Figure 4.17 Least Retweetable Words
in normal tweets would seem to predict (Figure 4.17). This list of
words is far less interesting than the list of most retweetable words,
and that’s the point: they’re boring.
Most of these words indicate that the person using them is talk-
ing about himself or herself and personal activities, such as watching
the game, listening to something, or going to bed. Even worse is the
occurrence of the word bored here. If you’re tweeting that you’re bored,
don’t expect it to get retweets, as you’re being quite boring yourself.
An analysis I did of 2.7 million link-containing tweets revealed
an interesting pattern that has implications for all kinds of commu-
nications professionals working with Twitter. I looked at each tweet
in my data set and identified those that were clicked on 0 times but
were retweeted at least once. I also identified those tweets that were
clicked on but were retweeted more times than they were clicked.
I found that 14.64 percent of the tweets in my study were never
clicked on but were retweeted and 16.12 percent of the tweets in
my database had more retweets than clicks (Figure 4.18). This tells
me that many people who will retweet an article will do so without
reading it first.
Although the sociological implications of these data could cer-
tainly be quite interesting, I’m mostly interested in what this means
for marketers. And what it means for marketers is that your head-
line is the most important piece of your content when it comes to
21. Twitter 69
14.64% of the retweeted
tweets had 0 clicks.
16.12% had more
retweets than clicks.
Figure 4.18 Retweets versus Clicks
Twitter success. If your headline doesn’t entice and motivate the
retweet, the body of your content might not ever get the chance
to succeed.
Even if much retweeting happens in the absence of clicks,
I’m still interested in getting people to click on the links I post to
Twitter. When it comes time to leverage the huge reach and retweet
counts we’ve built on Twitter into actual dollars-and-cents return
on investment, it’s all about how much traffic we can send to our
website and then convert into leads or customers.
When analyzing clicks on Twitter, I use a metric called click-
through rate (CTR). Marketers will be familiar with this from e-mail
marketing or pay per click (PPC), but it functions a little differently
on Twitter. I divide the number of clicks on a link by the number of
followers the user had when they sent the tweet in question.
The first thing I looked at when I began to study Twitter CTRs
was the length of the tweet, in characters. I found that longer tweets
(up to about 130 characters) tend to get more clicks than shorter
tweets (Figure 4.19).
Then I looked at the actual position within the tweets occupied
by links and its relationship to CTR. Most Twitter users, myself
included, typically put the link at the end of the tweet. The format
is generally: “Title of the content: http://linktothecontent.com.”
22. 70 Channels
18%
16%
14%
12%
10%
CTR
8%
6%
4%
2%
0%
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
105
110
115
120
125
130
135
140
Tweet Length
Figure 4.19 Longer Tweets Get More Clicks
200,000 Link-Containing Tweets Analyzed
Beginning Middle End
Figure 4.20 Twitter CTR Heat Map
To visualize this data, I created the heat map you see in
Figure 4.20. Each vertical bar represents a position of a link inside
of a tweet. Bars to the left represent links placed at the beginning of
the tweet; bars to the right indicate links placed at the end. The
darkness of the bar represents the average CTR of the links at each
position: the darker the bar, the higher the CTR.
My findings were surprising. Although there is a single dark bar
at the end of the heat map, there is a much larger sweet spot of clicks
about a quarter of the way into the tweet. After I first published
23. Twitter 71
these data, several people have told me that they’ve experimented
with a format like “new post: http://linktopost.com title of the post”
and it’s worked for them.
I’m not sure why this format works so well; perhaps it’s because
most Twitter accounts are still putting links at the end of tweets,
so tweets like this stand out. Experiment with it and see if it works
with your audience.
I also analyzed the CTRs of 20 highly followed Twitter
accounts, including mainstream news sources, such as the
Washington Post and the New York Times; geekier sources, such
as Mashable and Gizmodo; and celebrity accounts, such as the
Kardashians and Alyssa Milano. I found that there is no standard
CTR, as they vary widely across these accounts. The New York
Times has a very low CTR, whereas Alyssa Milano has a very
high CTR.
However, I did find one pattern that held true across all of
the accounts I looked at. When one of them tweet a link and
didn’t tweet another link for an hour, they had a certain CTR
(Figure 4.21). When they tweeted two links in an hour, the
CTR dropped. When it was three links in an hour, the CTR was
even lower. As the pace of link-tweeting increased, the CTR for
each link decreased.
300%
200%
CTR
100%
0%
1
2
3
4
5
6
7
8
9
10
11
12
13
Links Tweeted per Hour
Figure 4.21 CTR by Links Tweeted per Hour
24. 72 Channels
If you’re sharing content you’ve found from other sources across
the Web, tweet it as fast as you want. But when you’re tweeting your
own content to send traffic to your site, slow down and tweet it at a
more deliberate pace.
Another CTR analysis I did was on parts of speech. I analyzed
the four major parts of speech (adverbs, adjectives, nouns, and
verbs) and their relationship to the CTR of the tweets they were
found in. I compared the CTR of the individual tweets to the
average for the Twitter user to account for the wild differences
in CTRs.
I found that tweets heavy with adverbs and verbs performed
better than tweets with more nouns and adjectives (Figure 4.22).
Action-based words got more clicks than entity-based words. Action-
based words include calls to action, which is likely the cause of some
of this effect.
Don’t forget to experiment with action-based calls to action on
Twitter. If you want more clicks, ask for them. But be creative; don’t
just try “Click here,” but instead try “Check this out” or “Tell me
what you think.”
Use action words:
more verbs, fewer nouns.
After analyzing 200,000 link containing tweets, I found taht
tweets that contained more adverbs and verbs had
higher CTRs than noun- and adjective-heavy tweets.
3%
2%
Adjectives
Nouns
1%
Effect on CTR
0%
Adverbs
Verbs
–1%
–2%
–3%
–4%
–5%
Figure 4.22 Relationship between Various Parts of Speech and CTR
25. Twitter 73
Tweet later in the day.
I found that tweets posted in the afternoon hours
had higher CTRs than twees posted in the morning.
1.5%
1.0%
0.5%
Effect on CTR
0%
–0.5%
–1.0%
–1.5%
–2.0%
–2.5%
12 am
1 am
2 am
3 am
4 am
5 am
6 am
7 am
8 am
9 am
10 am
11 am
12 pm
1 pm
2 pm
3 pm
4 pm
5 pm
6 pm
7 pm
8 pm
9 pm
10 pm
11 pm
Figure 4.23 Relationship between Time of Tweet and CTR
I also looked at the hour of data and its relationship to CTRs and
found a pattern familiar if you think back to my data on retweeting
and time of day.
Tweets posted later in the day—afternoon Eastern time—
tended to have higher CTRs than tweets posted early in the morn-
ing (Figure 4.23). As with all timing data, be careful to test and
experiment with these findings, as your audience may behave dif-
ferently than the average of a very large data set.
And when I looked at day of the week and its relationship to
CTR, I found something surprising. Tweets that were posted toward
the end of the week got more clicks than those posted on Monday
through Wednesday (Figure 4.24). But it wasn’t just Thursday and
Friday that performed well; Saturday and Sunday also both have
high CTRs.
I follow thousands of accounts on Twitter. My Twitter stream
is very active during the business day Monday through Friday. On
the weekends, it moves much slower, and what content does come
26. 74 Channels
0.2%
0%
–0.2%
–0.4%
–0.6%
Mon Tue Wed Thu Fri Sat Sun
Figure 4.24 Effect of Day of the Week on CTR
through is often about sports and other non-work-related topics.
The few times something interesting about marketing does show up
on a Saturday or Sunday, it gets more of my attention because there
are fewer other things fighting for it. I call this contra-competitive
timing, and we’ll see examples of it throughout this book.
Don’t take these data to mean that you should tweet links only
on the weekends. Instead, take it as an invitation to experiment
with the weekends if you hadn’t already been using them.
27. 61 how to attract customers with twitter & Vine
measure THE
ROI OF TWITTER
Find out whether your investment in Twitter is paying off.
Sign up for HubSpot’s 30-day trial to start tracking some
insightful marketing analytics.
http://offers.hubspot.com/free-trial
Share This Ebook!
www.Hubspot.com
28. Available April 2013 wherever books and eBooks
are sold.
Pre-Order today from your favorite retailer!
http://www.wiley.com/buy/9781118138274